Synthesis and structure-activity relationships of phenyl-substituted coumarins with anti-tubercular activity that target FadD32

Article abstract of DOI:10.1016/j.bmcl.2013.09.035

In an effort to develop new and potent agents for therapy against tuberculosis, a high-throughput screen was performed against Mycobacterium tuberculosis strain H37Rv. Two 6-aryl-5,7-dimethyl-4-phenylcoumarin compounds 1a and 1b were found with modest activity. A series of coumarin derivatives were synthesized to improve potency and to investigate the structure-activity relationship of the series. Among them, compounds 1o and 2d showed improved activity with IC₉₀ of 2 μM and 0.5 μM, respectively. Further optimization provided compound 3b with better physiochemical properties with IC₉₀ 0.4 μM which had activity in a mouse model of infection. The role of the conformation of the 4- and 6-aryl substituents is also described.

Full text of DOI:10.1016/j.bmcl.2013.09.035

Bioorganic & Medicinal Chemistry Letters 23 (2013) 6052–6059
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and structure–activity relationships of phenyl-substituted
coumarins with anti-tubercular activity that target FadD32
Tomohiko Kawate a,b,c,d, , Noriaki Iwase a, , Motohisa Shimizu a, , Sarah A. Stanley a,b,c,d
,
Samantha Wellington a,d, Edward Kazyanskaya a, Deborah T. Hung a,b,c,d,
⇑
a The Broad Institute of MIT and Harvard, 7 Cambridge Center, Cambridge, MA 02142, USA
b Department of Molecular Biology, Massachusetts General Hospital, 185 Cambridge Street, Boston, MA 02114, USA
c Center for Computational and Integrative, Massachusetts General Hospital, 185 Cambridge Street, Boston, MA 02114, USA
d Department of Microbiology and Immunobiology, Harvard Medical School, 200 Longwood Avenue, Boston, MA 02115, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 16 July 2013
Revised 7 September 2013
Accepted 11 September 2013
Available online 19 September 2013
In an effort to develop new and potent agents for therapy against tuberculosis, a high-throughput screen
was performed against Mycobacterium tuberculosis strain H37Rv. Two 6-aryl-5,7-dimethyl-4-phenyl-
coumarin compounds 1a and 1b were found with modest activity. A series of coumarin derivatives were
synthesized to improve potency and to investigate the structure–activity relationship of the series.
Among them, compounds 1o and 2d showed improved activity with IC90 of 2
tively. Further optimization provided compound 3b with better physiochemical properties with IC90
0.4 M which had activity in a mouse model of infection. The role of the conformation of the 4- and 6-
aryl substituents is also described.
lM and 0.5 lM, respec-
Keywords:
Mycobacterium tuberculosis
Coumarin
Structure–activity relationship (SAR)
Inhibitor
l
Ó 2013 Elsevier Ltd. All rights reserved.
Tuberculosis (TB) is an infectious disease caused by the bacillus
Mycobacterium tuberculosis (Mtb). It is estimated that one-third of
the world’s population is infected by Mtb in its latent form and that
there were 9.4 million new TB cases and 1.7 million deaths from
active TB in 2009, according to the World Health Organization.1
Thus, despite the availability of effective anti-TB drugs, such as iso-
niazid and rifampin, TB is still one of the world’s leading causes of
deaths by an infectious disease. Development of new agents that
reduce the duration and complexity of the current therapeutic reg-
imen, as well as effectively treat multi-drug resistant (MDR) and
extensively drug resistant (XDR) TB, would have a major impact
on public health.
role in the biosynthesis of the unique branched fatty acids (mycolic
acids) that make up the Mtb cell wall.4 This is a known validated
pathway for TB therapeutics as demonstrated by isoniazid, a lynch-
pin of TB therapy that also targets mycolic acid biosynthesis by
inhibiting the enzyme InhA. However, increasing isoniazid resis-
tance may soon limit the efficacy of this drug and thus new candi-
dates to replace isoniazid are urgently needed. Targeting
a
different enzyme, such as FadD32, in the same biosynthetic path-
way is one way to circumvent isoniazid resistance while inhibiting
a well-validated function. We now report studies defining the
structure activity relationship of a series of 4,6-bis-aryl substituted
5,7-dimethyl-coumarins with anti-mycobacterial activity as well
as the effect of the twisted conformation of these aryl groups.
Of the compounds identified in the primary screen, replacement
of the methyl group of compound 1a with hydroxylmethyl (1b) re-
sulted in improved activity. Therefore, we first sought to optimize
the substituent at the 6-position (Table 1). Synthesis of 6-substi-
tuted 5,7-dimethyl-4-phenyl-2H-chromen-2-one 1a, 1b, 1f–1s,
and 1u–1x, was achieved by following a literature procedure5 with
small modifications as shown in Scheme 1. Namely, the key inter-
mediate bromide, 6-bromo-5,7-dimethyl-4-phenyl-2H-chromen-
2-one 8, was prepared from phenyl-propiolate ester 7 by Pd(OAc)2
catalyzed cyclization.5 Using a Suzuki–Miyaura cross-coupling
(SMC) reaction of bromide 8 with the corresponding boronic acids,
1a, 1b, 1f–1s and 1u–1x were synthesized in the presence of
2-dicyclohexylphosphino-20-methylbiphenyl and Pd(OAc)2 or
Toward that end, we performed a chemical high-throughput
screen to identify inhibitors of growth of GFP expressing Mtb
(H37Rv), screening 20,000 compounds at the Broad Institute. Four
compounds 1a–1d (IC90 36 lM, 7 lM, 184 lM and 295 lM, respec-
tively, Fig. 1) were identified, thus defining a novel class of com-
pounds containing the coumarin scaffold with anti-tubercular
activity.2 We have previously reported that this coumarin class tar-
gets the fatty acyl ACP synthetase activity of the FadD32 enzyme.3
This enzyme is essential for Mtb survival because it plays a critical
⇑
Corresponding author.
E-mail addresses: hung@molbio.mgh.harvard.edu, hung@broadinstitute.org (D.T.
Hung).
These authors contributed equally to this work.
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.09.035

T. Kawate et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6052–6059
6053
The anti-tubercular activity of the synthesized analogues (1a,
1b, 1e–1x and 8) was evaluated against Mtb strain H37Rv using
a 96 well plate based assay in which the growth inhibitory activity
of each analog was evaluated using optical density at 600 nm
(OD600) as a readout. Mtb was exposed to inhibitor for a period
of 7 days at which time the IC90 was determined (Table 1).3 Since
a hydro- or bromo- substitution at the 6-position of the coumarin
core, corresponding to 1e and 8, respectively, essentially abolished
all activity, we focused on evaluating substitutions on the 6-phenyl
group. The electron withdrawing groups (1f–1l) substituted at the
para (1f, 1h, 1j) and ortho (1l) positions preserved activity, while
substitution at the meta position (1i, 1k) reduced activity, with
the exception of the small fluorine group (1g). In the case of elec-
tron donating group substitutions, a methoxy group (1m, 1n) at
the ortho position (1n) abolished activity while meta substitution
Figure 1. Structures of HTS hit coumarins.
[1,10-Bis(diphenylphosphino)ferrocene]dichloropalladium(II)
dichloromethane complex (PdCl2(dppf)ꢀCH2Cl2). 6-Hydro com-
pound 1e was isolated as a side-product of the SMC reaction and
racemic 1t was synthesized by sodium borohydride reduction of
acetyl compound 1s (see Supplementary data).
Table 1
Structures and anti M. tuberculosis activity of coumarins 1a–1b, 1e–1x and 8
Compda
R
Yieldb (%)
ClogP13
IC90 (lM)
1a (CCA1)a
4-Me-Ph
4-Hydroxymethyl-Ph–
H
4-F-Ph
3-F-Ph
4-CF3-Ph
3-CF3-Ph
4-Cl-Ph
3-Cl-Ph
2-Cl-Ph
3-MeO-Ph
2-MeO-Ph
4-Pyridyl
87c
81c
38e
76c
91c
96c
87c
29c
78c
21d
46c
19c
89c
79c
31c
69d
46c
61f
27c
32d
16d
41c
-
6.085
4.548
4.298
5.729
5.729
6.469
6.469
6.299
6.299
6.049
5.505
4.945
4.089
4.089
3.132
4.288
5.025
4.857
5.055
7.259
4.747
7.489
5.161
36
7
1100
26
21
29
No killing at 64
54
260
22
1b* (CCA2)a
1e
1f
1g
1h
1i
1j
l
M
M
1k
1l*
1m
31
1n
No killing at 64
2
69
8
36
220
4
7
l
1o* (CCA26)a
1p
1q
1r
3-Pyridyl
5-Pyrimidyl
2-Me-4-pyridyl
4-Acetyl-Ph
4-(1-Hydroxyethyl)Ph
4-(Methylcarbamoyl)Ph
4-Me-1-naphthyl
4-Hydroxymethyl-2-Me-Ph
2-(p-Tolyl)ethynyl
Br
1s
1t*
1u*
1v
1w*
1x
8
No killing at 64
22
61
l
M
M
No killing at 64
l
a
b
c
d
e
f
Compound number from previous annotation.3
Isolated yield of Suzuki–Miyaura coupling reaction.
PdCl2(dppf)ꢀCH2Cl2 was used as catalyst.
Pd(OAc)2-2-dicyclohexylphosphino-20-methylbiphenyl was used as catalyst.
Isolated as side product.
Yield of NaBH4 reduction of 1s.
*
Analogue tested against FadD32 mutant (FadD32E120G) and for toxicity against HepG2 cells.
a
b
c
Br
Br
R
+
Br
HO
O
O
O
O
O
OH
O
O
5
6
7
8
1a, 1b, 1e-1s, 1u~1x
Scheme 1. Synthesis of compounds 1a, 1b, 1e–1s, 1u–1x. R is as defined in Table 1. Reagents and conditions: (a) DCC, DMAP, CH2Cl2, 100%; (b) Pd(OAc)2, TFA, CH2Cl2, 77%; (c)
R-B(OH)2, catalyst (PdCl2(dppf)CH2Cl2 or 2-dicyclohexylphosphino-20-mehtylbiphenyl/Pd(OAc)2), base (K2CO3 or K3PO4), 1,4-dioxane, H2O, 16-93%.

6054
T. Kawate et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6052–6059
(1m) showed similar activity to 1a. Replacement of the phenyl
group with other heteroaryl groups identified the 4-pyridyl com-
pound 1o as being the most potent of this series.
In testing the activity of these morpholine coumarin analogues
(Table 3), we found that changing the phenyl group to a 3-anilino
(3b) or 3-pyridyl (3d) group were well-tolerated. However, a 4-ani-
lino group (3a), 3-nitrophenyl group (3c), 4-methoxyphenyl group
(3h) and 3-methoxycarbonylphenyl group (3i) at the 4-position all
resulted in reduced activity. Additionally, the N-substituted 3-ani-
lino compounds (3e, 3f, 3g) were less potent than the parent com-
pound (3b), indicating that there is a size limit for substitution at
the 4-position of coumarin core or a preference for the primary
amine.
Finally, we turned to examine how the activity might be af-
fected by the conformational constraints placed on the 4,6-diaryl
substituents by the 5,7-dimethyl substitution pattern. We hypoth-
esized that the methyl groups at the 5- and 7-positions should re-
strict rotation of aryl ring at both the 4- and 6-positions. Torsion
angle histograms derived from the Cambridge Structural Database
(CSD)7,8 showed that the substituent at the 5-position had rela-
tively little impact on the torsion angle of the 4-phenyl group rel-
ative to the coumarin ring. In contrast, the torsion angle profiles of
the 6-aryl group relative to the coumarin ring were significantly af-
fected by the substitution patterns at the 5- and 7-positions
(Fig. 2). The CSD predicted torsion angles of both aryl groups at
4- and 6-position was confirmed in the crystal structure of 1b9
(Figs. 2 and 3). The crystal structure also revealed a subtle bending
of the phenyl group at the 4-position away from the 5-methyl
group, as well as a slight twist of the pyrone ring in coumarin core.
These small effects likely explain the small difference in 4-aryl ring
torsion angles observed compared to the CSD histograms.
To examine the importance of the presence of 4-phenyl group
and torsion angle of the aryl substituent at 6-position for their
anti-tubercular activity, we systematically removed phenyl or
methyl substituents and measured their impact on activity (Ta-
ble 4). We synthesized two analogues of compound 1a (Scheme 5)
and four analogues of compound 1b (Schemes 6 and 7). Synthesis
of compounds 4a, where the 4-phenyl substituent has been re-
placed by methyl, and 4b, where the 5,7-dimethyl groups have
been replaced by hydrogen, is shown in Scheme 5. The substituted
4-bromophenols 21a and 21b were converted to the biphenyl
intermediates 22a and 22b by SMC reaction with p-tolylboronic
acid. Esterification of 22a and 22b with substituted propiolic acid
gave the corresponding esters 23a and 23b followed by Pd(OAc)2
catalyzed cyclization5 to give the products 4a and 4b.
These results suggest that hydrophilic or basic groups on the
molecular surface improved activity. Therefore, we synthesized a
series of substituted (aminomethyl)phenyl analogues to improve
potency (Table 2). Synthesis of 6-(4-(substituted amino)methyl)-
phenyl-5,7-dimethyl-4-phenyl-2H-chromen-2-one 2a–2e is de-
picted in Scheme 2. In the presence of corresponding amine,
PdCl2(dppf)ꢀCH2Cl2 catalyzed SMC reaction of bromide 8 and 4-
(bromomethyl)phenylboronic acid yielded the desired substituted
amino analogues of 1a (2a–2e).
All of the analogues of this series showed good activity (Table 2).
While the bulky hydrophobic substitution (cyclohexyl 2b) and long
chain hydrophobic substitution (n-hexyl 2c) were tolerated, the
morpholino substitution (2d) gave a >10-fold increase in potency
relative to the original screening hits. From these data, it appears
that the hydrophilicity of the molecular surface at this position is
important.
With the 6-position fixed as a 4-(4-morpholino)methylphenyl
group, we turned to explore alterations to the 4-position of cou-
marin core to improve both physical properties and potency by
introducing polar or charged groups (Table 3). Synthesis of 4-
substituted-5,7-dimethyl-6-(4-(morpholinomethyl)phenyl)-2H-
chromen-2-one 3a–3g, analogues of 2d, is shown in Scheme 3.
Propiolic acid 11 was first converted to ester 12 followed by a
modified Sonogashira reaction6 with the corresponding iodides
to give the intermediates 13a–13c. Cyclization5 of 13 in the pres-
ence of Pd(OAc)2 catalyst gave the bromides 14a–14c. Reduction of
the nitro substituent on 14a and 14b with Fe yielded aniline 15a
and 15b, respectively. The bromides 15a,b and 14b,c were con-
verted to the morpholino compounds 3a–3d by the same condi-
tions described above. The aniline 3b was further converted to
analogues 3e–3g by reacting with the corresponding acid
chlorides.
Compounds 3h and 3i were obtained through a different route,
shown in Scheme 4. First, a modified Sonogashira reaction6 of tert-
butyl propiolate 16 with the corresponding iodides resulted in
intermediates 17a and 17b, which were then treated with trifluo-
roacetic acid or formic acid to yield the carboxylic acids 18a and
18b. The carboxylic acid intermediates were coupled with 4-bro-
mo-3,5-dimethylphenol
using
N,N0-dicyclohecylcarbodiimide
(DCC) to give ester 19a and 19b. Cyclization5 of 19 in the presence
of Pd(OAc)2 catalyst gave bromides 20a and 20b, followed by SMC
reaction with 4-(bromomethyl)phenylboronic acid in the presence
of morpholine to give the products 3h and 3i.
Synthesis of 5,7-substituted 6-(4-(hydroxymethyl)phenyl)-4-
phenyl-2H-chromen-2-one 4c–4e, which systematically interro-
gate the need for the 5,7-methyl groups for activity, is shown in
Scheme 6. SMC reaction of the substituted 4-bromophenols 24a
Table 2
SAR of compounds 2a–2e
R1
N
R2
O
O
Compd
R1
R2
Yieldb (%)
ClogP13
IC90
19
5–16
4
0.5
5.894 6
(lM)
2a*
–CH2CH2CH2CH2CH2–
Cyclohexyl
n-Hexyl
–CH2CH2OCH2CH2–
–CH2CH2N(Me)CH2CH2–
81
60
33
89
6.613
6.985
7.730
5.333
20
2b*
H
H
2c
2d* (CCA31)a
2e*
a
Compound number from previous annotation.3
Isolated yield of Suzuki–Miyaura coupling reaction.
b
*
Analogue tested against FadD32 mutant (FadD32E120G) and for toxicity against HepG2 cells.

T. Kawate et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6052–6059
6055
R1R2NH
PdCl2(dppf)•CH2Cl2
K2CO3
R1
Br
N
+
R2
Br
(HO)2B
1,4-Dioxane
20~89%
10
O
O
O
O
2a-2e
8
Scheme 2. Synthesis of coumarins 2a–2e. R1 and R2 are defined in Table 2.
Table 3
SAR of compounds 3a–3i
Finally, we synthesized a structural variant in which we moved
the 5,7-dimethyl coumarin substituents onto the ortho positions of
the 6-phenyl ring, which would be predicted to create a similar
torsion angle between the 6-phenyl ring and the coumarin core.
Synthesis of 6-(4-(hydroxymethyl)-2,6-dimethylphenyl)-4-phe-
nyl-2H-chromen-2-one 4f is shown in Scheme 7. Starting from
2-bromo-1,3,5-trimethylbenzene 29, TBS protected (4-bromo-3,5-
dimethylphenyl)methanol 32 was obtained through CrO3 oxida-
tion12, NaBH4 reduction and TBS protection. SMC reaction of
bromide 32 with 4-(hydroxy)phenylboronic acid resulted the
phenol 33, which was converted to 4f through esterification,
cyclization5 and deprotection.
The anti-tubercular activity of 4a–4f is summarized in Table 4.
The loss of activity of the 4-methyl compound 4a indicates the
importance of the presence of 4-phenyl group, as was similarly
observed with the loss of the 6-phenyl group in 1e. Both ortho
non-substituted analogues at the 5- and 7-positions (4b and 4c)
were either inactive or had at least 10 times less activity than
the di-substituted counter part (1a or 1b). The mono-substituted
compounds (4d and 4e) had intermediate activity. Loss of either
methyl group results in the same IC90, suggesting that it is not
any effect on the 4-phenyl but rather their effect on the 6-phenyl
ring that is likely important. Further, though they may play an
additional role in target site interactions, the equal loss of activity
given their symmetry relative to the 6-phenyl group suggests that
one important role of the methyl groups may be to constrain the
torsion angle of the 6-phenyl group. These findings reinforce the
likely importance of the torsion angle of 6-aryl group for anti-
mycobacterial activity, with the optimal torsion angle favored by
the dimethyl substitution.
Compd
R
ClogP13
IC90 (lM)
2d* (CCA31)a
Ph
5.333
4.327
4.327
5.076
3.902
4.555
4.297
5.497
5.363
0.5
14
0.4
6
3a*
4-NH2-Ph
3-NH2-Ph
3-NO2-Ph
3-Pyridyl
3b (CCA34)a
3c*
3d*
3e*
3f
1
3-Acetamido-Ph
12
59
39
3-(Methylsulfoamido)-Ph
3-((Ethoxycarbonyl)amino)-Ph
4-MeO-Ph
3g
3h
No killing at
62
59
lM
3i
3-(Methoxycarbonyl)-Ph
5.302
a
Compound number from previous annotation.3
*
Analogue tested against FadD32 mutant (FadD32E120G) and for toxicity against
HepG2 cells.
and 24b with 4-(hydroxymethyl)phenylboronic acid in the pres-
ence of PdCl2(dppf)ꢀCH2Cl2 gave the diols 25a and 25b. Protection
of the diol 25a and 25b with tert-Butyldimethylsilyl chloride
(TBSCl) gave the diTBS intermediates 26a and 26b followed by
selective deprotection11 with 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU) to give the phenols 27a and 27b. Esterification with 3-phe-
nylpropiolic acid, followed by Pd(OAc)2 catalyzed cyclization5 and
deprotection gave the coumarins 4c–4e. The isomers 4d and 4e
were separated by preparative HPLC.
Finally, we synthesized 4f, moving the 5,7-dimethyl groups
onto the ortho positions of the 6-pheny ring, which is predicted
H
H
Ar
Ar
Br
Br
a
b
c
Br
Br
HO
O
O
O
OH
O
O
O
O
11
6
12
13
14, 15
14a; Ar=4-NO2-Ph
13a; Ar=4-NO2-Ph
13b; Ar=3-NO2-Ph
13c; Ar=3-pyridyl
14b; Ar=3--NO2-Ph
d
15a; Ar=4-NH2-Ph
15b; Ar=3-NH2-Ph
14c; Ar=3-pyridyl
NHR
Ar
O
N
N
f
e
O
O
3b
O
O
O
3e-3g
3a-3d
Scheme 3. Synthesis of compounds 3a–3g. Reagents and conditions: (a) DCC, DMAP, CH2Cl2, 51%; (b) Ar-I, Pd(PPh3)2Cl2, CuI, K2CO3, THF, 27–74%; [3a, 3b] (c and d) Pd(OAc)2,
TFA, CH2Cl2 then Fe, AcOH, EtOH, 0.8–16%; (e) morpholine, 4-(bromomethyl)phenylboronic acid, PdCl2(dppf)ꢀCH2Cl2, K2CO3, 1,4-dioxane, H2O, 43–75%; [3d] (c and e)
Pd(OAc)2, TFA, CH2Cl2 then morpholine, (4-(bromomethyl)phenyl)boronic acid, PdCl2(dppf)ꢀCH2Cl2, K2CO3, 1,4-dioxane, H2O, 1.3%; (f) R-Cl, Et3N or pyridine, CH2Cl2, 50–100%.

6056
T. Kawate et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6052–6059
R
R
a
b
c
O
O
O
O
O
OH
18
16
17
17a: R=4-MeOPh
18a: R=4-MeOPh
17b: R=3-MeOC(O)Ph
18b: R=3-MeOC(O)Ph
R
R
R
N
d
e
Br
O
Br
O
O
O
O
O
O
19
20
3h, 3i
20a: R=4-MeOPh
19a: R=4-MeOPh
3h: R=4-MeOPh
20b: R=3-MeOC(O)Ph
19b: R=3-MeOC(O)Ph
3i: R=3-MeOC(O)Ph
Scheme 4. Synthesis of compounds 3h and 3i. Reagents and conditions: (a) R-I, PdCl2(Ph3P)2 or Pd(PPh3)4, CuI or ZnBr2, K2CO3 or LDA, THF, 3–99%; (b) TFA/CH2Cl2 or formic
acid, 70–94%; (c) 4-bromo-3,5-dimethylphenol, DCC, DMAP, CH2Cl2, 49–62%; (d) Pd(OAc)2, TFA, CH2Cl2, 4–73%; (e) morpholine, 4-(bromomethyl)phenylboronic acid,
PdCl2(dppf)ꢀCH2Cl2, K2CO3, 1,4-dioxane, H2O, 53–54%.
Figure 2. Torsion angle histograms and search structures. Torsion histograms for the biaryl ring and their search structures in the CSD8 (Q = C, N, O). (a) Naphthyl–phenyl
type biaryl models for 5-hydro- (top) and 5-methyl- (bottom) coumarins, (b) biaryl derivatives with no ortho substituents (top), with one ortho substituent (middle) and with
two ortho substituents (bottom). Red bars indicate torsion angles observed in the crystal structure of 1b (Fig. 3).
to have the same active torsion angle as the original 1b. 4f was less
potent with an IC90 170 M compared to 7 M for 1b. This could be
to an electron rich pi-stacking interaction between FadD32 and the
phenyl ring of the coumarin. Clearly a crystal structure of these
4,6-diaryl coumarins bound to FadD32 in the future will help to
clarify these interactions. No obvious correlation was observed
between anti tubercular activity and ClogP value13 of coumarin
compounds 1a, 1b, 1e–x, 2a–e, 3a–i, 4a–f and 8 (Tables 1–4).
In order to confirm that synthesized analogues were still on
l
l
explained by steric hindrance of the methyl groups overlying the
phenyl ring of the coumarin core. Such reduction of activity was
also observed in ortho substituted analogs 1n and 1v, compared
to 1m and 1a, respectively. On the other hand, the potency of ortho
chloro analog 1l improved compared to 1a, 1j and 1k, suggests that
the interaction between molecule and FadD32 may be more com-
plex, possibly with the electron density of the halogen contributing
target against FadD32, we tested representative compounds (Ta-
3
ble 1–3) against a FadD32 mutant of M. tuberculosis (FadD32E120G
)

T. Kawate et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6052–6059
6057
Figure 3. Crystal structure of coumarin 1b (stereo view).10
Table 4
SAR of compounds 4a–4f
R1 R2
R3
R4
O
O
Compd
R1
R2
R3
R4
ClogP13
IC90
36
(lM)
1a
Ph
Me
Ph
Ph
Ph
Ph
Ph
Ph
Me
Me
H
Me
H
H
Me
H
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-HOCH2-Ph
4-HOCH2-Ph
4-HOCH2-Ph
4-HOCH2-Ph
Me
Me
H
Me
H
Me
H
H
6.085
4.696
5.687
4.548
4.150
4.349
4.349
4.548
4a (CCA42)a
4b (CCA44)a
1b
No killing at 250
lM
b
7
115
68
70
170
4c (CCA45)a
4d (CCA46)a
4e (CCA47)a
4f
2,5-DiMe-4-(HOCH2)-Ph
a
Compound number from previous annotation.3
b
24% Inhibition at 250
lM.
R1
R2
R1 R2
R2
R2
Br
b
c
a
R4
HO
R4
O
O
R4
HO
O
O
R4
21
22
23
4
4a: R1=R2=R4=Me
21a: R2=R4=Me
22a: R2=R4=Me
23a: R1=R2=R4=Me
4b: R1=Ph, R2=R4=H
21b: R2=R4=H
22b: R2=R4=H
23b: R1=Ph, R2=R4=H
Scheme 5. Synthesis of compounds 4a and 4b. Reagents and conditions: (a) p-tolylboronic acid, PdCl2(dppf)ꢀCH2Cl2, K2CO3, 1,4-dioxane-H2O (9:1), 70–100%; (b) R1-CC-COOH,
DCC, DMAP, CH2Cl2, 82–100%; (c) Pd(OAc)2, TFA, CH2Cl2, 33–57%.
that is resistant to 1b (CCA2) and 3b (CCA34). We confirmed their
activity at FadD32 by demonstrating that the FadD32 mutant is
also resistant to all synthesized, tested analogues, with none of
them having any activity on the mutant at the highest concentra-
Further, increased potency can be gained by increased basicity of
the 6-phenyl ring, with a para substituted morpolino resulting in
significant increase in potency (2d; IC90 0.5
4-phenyl ring did little to improve potency, though converting it
to the 3-aniline decreased the IC90 to 0.4 M (3b). It is important
lM). Changes on the
tion tested (125
compound 2b showed only a small shift in IC90 (from ꢁ16
the wild-type to 32 M in the FadD32 mutant) suggesting that
l
M), with one exception.3 Interestingly,
M in
l
l
to note that the aniline is not critical for activity in future develop-
ment efforts given concerns of its role as a pharmacophore in
drugs, as it only gives a small incremental increase in potency.
With the goal of demonstrating in vivo activity of this coumarin
series, we selected four compounds, including the most potent 2d
and 3b, for characterization of their physicochemical properties.
The results are summarized in Table 5. As predicted, introduction
of a morpholino group (2d) greatly increased water solubility
compared to the original hit 1b, and further improvement was ob-
served by adding the additional amino group as in the 3-anilino
compound 3b, though clearly the aniline is not necessary for activ-
ity. In addition, stability in human plasma was also improved in
l
2b may have acquired an alternative or additional target. We also
tested the representative compounds for toxicity against HepG2
cells and found that like the previous lack of host cell toxicity de-
scribed for 3b (CCA34),3 none of the synthesized, tested analogues
showed any toxicity against HepG2 cells at 125
exception of 2b, which had a LD50 32 M.
lM, again with the
l
To summarize, we found that in this coumarin series, both 4-
and 6-phenyl substituents are important for anti-mycobacterial
activity. The 5,7-methyl substituents are also important as they
likely help to constrain the torsion angle of the 6-phenyl group.

6058
T. Kawate et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6052–6059
R2
R2
OH
R2
X
R2
OTBS
d
e, f
Br
a
R4
R4
HO
R4
O
O
R4
O
O
Y
28
4c-4e
24
25: X=Y=OH
b
24a: R2=R4=H
28a: R2=R4=H
4c: R2=R4=H
26: X=Y=OTBS
27: X=OTBS,Y=OH
24b: R2=Me, R4=H
28b: R2=Me, R4=H
c
4d: R2=H, R4=Me
4e: R2=Me, R4=H
25a, 26a, 27a: R2=R4=H
25b, 26b, 27b: R2=Me, R4=H
Scheme 6. Synthesis of compounds 4c–4e. Reagents and conditions: (a) (4-(hydroxymethyl)phenyl)boronic acid, PdCl2(dppf)ꢀCH2Cl2, K2CO3, 1,4-dioxane-H2O (9:1), 42–86%;
(b) TBSCl, imidazole, CH2Cl2, 68–92%; (c) DBU, MeCN-H2O, 89–99%; (d) 5, DCC, DMAP, CH2Cl2, 46–83%; [4c] (e) Pd(OAc)2, TFA, CH2Cl2, then NaHCO3, 17%; [4d, 4e] (e) Pd(OAc)2,
TFA, CH2Cl2, then NaHCO3, (f) preparative HPLC, 3.2–4.0%.
OTBS
OAc
OAc
OH
c
a
d
b
Br
Br
Br
Br
29
30
31
32
OTBS
OH
OH
e
f
HO
O
O
O
O
33
34
4f
Scheme 7. Synthesis of compound 4f. Reagents and conditions: (a) CrO3, AcOH, Ac2O, cH2SO4, 14%; (b) NaBH4, THF, MeOH, H2O, 99%; (c) TBSCl, imidazole, CH2Cl2, 81%; (d) (4-
hydroxyphenyl)boronic acid, PdCl2(dppf)ꢀCH2Cl2, K2CO3, 1,4-dioxane–H2O (9:1), 38%; (e) 5, DCC, DMAP, CH2Cl2; (f) Pd(OAc)2, TFA, CH2Cl2, then NaHCO3, 2.2% (2steps).
Table 5
Physicochemical properties of selected compoundsa,3
Compd
Mol. wt.
Solubility (water,
lM)
Protein
Plasma
Protein bindingb (%bound)
Plasma stabilityc (%remaining)
1b
1o
2d
3b
356.41
327.38
425.52
440.53
0.2
3.2
36.1
89.3
99.4
98.8
98.6
98.7
80.3
83.5
84.4
88.0
a
b
c
Quantified by UPLC–MS.
Human plasma, after 5 h.
After 5 h in human plasma.
Table 6
Selected ADME and PK properties of compound 3ba,b
In vitro assay
MLMc
Intrinsic CL
(ml/min/kg)
MLMc
Half-life
(min)
Mouse plasma
Stability
(remaining at 8 h)
Caco-2 permeability (ꢂ10ꢃ6 cm/s)
A to B
B to A
80.33
8.63
111.5%
Vde (L/kg)
3.91
34.46
76.87
In vivo assayd
Half-life (h)
1.45
Clearance (ml/min/kg)
31.08
Bioavailability (%F)
61.95
a
b
c
Dihydrochloride salt (3bꢀ2HCl) was used for measurement.
Quantified by LC–MS/MS.
Mouse liver microsome.
Male Balb/c mice were used.
Volume of distribution.
d
e
compound 3b. We also measured selected absorption distribution
metabolism and elimination (ADME) and pharmacokinetic (PK)
properties of 3b and the results are summarized in Table 6.
Importantly, the lead compound 3b has good oral bioavailability
(62%), which is an important feature of antitubercular antibiotics.
We have now reported their activity against not only drug

T. Kawate et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6052–6059
6059
sensitive M. tuberculosis, but also drug resistant M. tuberculosis and
References and notes
other pathogenic mycobacterial species including Mycobacterium
marinum and Mycobacterium avium complex, and we have previ-
ously shown that they have good efficacy against M. tuberculosis
and M. marinum in a mouse and zebrafish model of infection,
respectively.3
In conclusion, we have discovered a novel coumarin series that
effectively kill Mycobacterium tuberculosis as well as other patho-
genic mycobacterial species. Through a structure–activity relation-
ship study and physicochemical property optimization, we have
achieved about 20-fold increase of potency and 450-fold improve-
ment in water solubility from the HTS hit compound, which has
allowed us to both identify the mechanism of action, targeting
FadD32, and to demonstrate good activity in two animal models
in our prior studies.3 In addition, the lead compound has excellent
oral bioavailability. Thus, in the setting of increasing resistance to
current anti-tubercular antibiotics, this class potentially defines a
promising new candidate series for development.
1. Global tuberculosis control: WHO report 2010. WHO//HTM/TB/2010.7.
2. Recently reported anti-tubercular compounds with coumarin core: (a)
Ananthan, S.; Faaleolea, E. R.; Goldman, R. C.; Hobrath, J. V.; Kwong, C. D.;
Laughon, B. E.; Maddry, J. A.; Mehta, A.; Rasmussen, L.; Reynolds, R. C.; Secrist, J.
A., III; Shindo, N.; Showe, D. N.; Sosa, M. I.; Suling, W. J.; White, E. L. Tuberculosis
2009, 89, 334; (b) Virsdoia, V.; Shaikh, M. S.; Manvar, A.; Desai, B.; Parecha, A.;
Loriya, R.; Dholariya, K.; Patel, G.; Vora, V.; Upadhyay, K.; Denish, K.; Shah, A.;
Coutinho, E. C. Chem. Biol. Drug Des. 2010, 76, 412; (c) Tandon, R.; Pannan, P.;
Aggarwal, N.; Pathak, R.; Baghel, A. S.; Gupta, G.; Arya, A.; Nath, M.; Parmar, V.
S.; Raj, H. G.; Prasad, A. K.; Bose, M. J. Antimicrob. Chemother. 2011, 66, 2543.
3. Stanley, S. A.; Kawate, T.; Iwase, N.; Shimizu, M.; Clatworthy, A.; Kazyanskaya,
E.; Sacchettini, J. C.; Ioerger, T. R.; Siddiqi, N.; Minami, S.; Aquadro, J. A.; Grant,
S. S.; Rubin, E. J.; Hung, D. T. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 11565.
4. (a) Portevin, D.; de Sousa-D’Auria, C.; Montrozier, H.; Houssin, C.; Stella, A.;
Lanéelle, M.-A.; Bardou, F.; Guilhot, C.; Daffé, M. J. Biol. Chem. 2005, 280, 8862;
(b) Léger, M.; Gavalda, S.; Guillet, V.; Rest, B. v. d.; Slama, N.; Montrozier, H.;
Mourey, L.; Quémard, A.; Daffé, M.; Marrakchi, H. Chem. Biol. 2009, 16, 510; (c)
Trivedi, O. A.; Arora, P.; Sridharan, V.; Tickoo, R.; Mohanty, D.; Gokhale, R. S.
Nature 2004, 428, 441; (d) Takayama, K.; Wang, C.; Besra, G. S. Clin. Microbiol.
Rev. 2005, 18, 81; (e) Carroll, P.; Faray-Kele, M.-C.; Parish, T. Appl. Environ.
Microbiol. 2011, 77, 5040.
5. (a) Li, K.; Zeng, Y.; Neuenswander, B.; Tunge, J. A. J. Org. Chem. 2005, 70, 6515;
(b) Jia, C.; Piao, D.; Oyamada, J.; Lu, W.; Kitamura, T.; Fujiwara, Y. Science 1992,
2000, 287; (c) Jia, C.; Lu, W.; Oyamada, J.; Kitamura, T.; Matsuda, K.; Irie, M.;
Fujiwara, Y. J. Am. Chem. Soc. 2000, 122, 7252; (d) Jia, C.; Piao, D.; Kitamura, T.;
Fujiwara, Y. J. Org. Chem. 2000, 65, 7516.
6. (a) Eckert, T.; Ipaktschi, J. Synth. Commun. 1998, 28, 327; (b) Anastasia, L.;
Negishi, E. Org. Lett. 2001, 3, 3111.
7. (a) Allen, F. H. Acta Crystallogr. B 2002, B58, 380; (b) Brameld, K. A.; Kuhn, B.;
Reuter, D. C.; Stahl, M. J. Chem. Inf. Model. 2008, 48, 1.
Acknowledgments
We thank Dr. Stephen Johnston of Broad Institute for physico-
chemical property measurements, Dr. Michael Serrano-Wu of
Broad Institute for coordinating ADME-PK analysis and Dr. Shao-
Ling Zheng of Harvard University for his help on X-ray data
collection and structure determination.
8. CSD search: CSD version 5.32 (November 2010) + update (Nov 2010). All search
were performed for structures with R factor <=0.05 (not disorderd, no errors,
not polymeric).
9. Crystal structure is deposited at Cambridge Crystallographyc Data Centre with
deposition number CCDC 94074 and reported in Ref. 3.
10. The stereo view of 1b was created by Olex2. Dolomanov, O. V.; Bourhis, L. J.;
Gildea, R. J.; Howard, J. A. K.; Puschmann, H. J. Appl. Crystallogr. 2009, 42, 339.
11. Yeom, C.; Kim, H. W.; Lee, S. Y.; Kim, B. M. Synlett 2007, 146.
12. Ayer, W. A.; McCaskill, R. H. Can. J. Chem. 1981, 59, 2159.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2013.
09.035.
13. CLogP values were calculated with ChemDraw Ultra v12.0.3, Perkin-Elmer.